In this guide, I will show you how to perform data normality tests in GraphPad Prism. Prior to any statistical analysis, the first thing you should do is to see whether your data is normally distributed. There are three normality tests in GraphPad Prism: D’Agostino-Pearson omnibus, Shapiro-Wilk and Kolmogorov-Smirnov tests. These will reveal more about the dataset and ultimately help in deciding which statistical test you should perform.
How to test for normality in GraphPad Prism
For this example, I will use the following dataset contained within a column table. I have two groups of data: males and females ages in years.
Which normality test should you perform?
GraphPad Prism allows for the option of three normality tests. So, which one should you go for? It is recommended to use the D’Agostino-Pearson omnibus test since it is easier to understand how it works. The Shapiro-Wilk test is useful when no two values are the same in the dataset. Finally, the general consensus is to avoid the use of the Kolmogorov-Smirnov test as it is now redundant.
I recommend reading the Q&A: Normality Tests sheet on the GraphPad website for more information section on normality tests
Performing the normality tests
Here is how you can perform normality tests in GraphPad Prism.
- In the data table view, click the ‘Analyze‘ button in the ‘Analysis‘ section of the ribbon at the top.
2. The ‘Analyze Data‘ window should now open. Click the ‘Column analyses‘ dropdown option, and under these options select ‘Column statistics‘. On the right-hand window, select the datasets which you want to analyze. Those with a tick will be included in the analysis. Then, click the ‘OK‘ button.
3. The ‘Parameters: Column Statistics‘ should now open. This window enables you to specify what to report in the analysis, and most importantly, what normality test(s) to perform. You can find these tests under the ‘Test if the values come from a Gaussian distribution‘ header. Select which test(s) you want to perform by clicking them and then click the ‘OK‘ button. For the purpose of this example, I will select all three tests.
A new results sheet should now open. The results of the normality tests should be reported near the bottom of this sheet.
- The test statistics (K2 or W or KS distance): These are the test statistics for the corresponding normality test.
- P value: The P value for each test.
- Passed normality test (alpha = 0.05)?: A simple ‘Yes’ or ‘No’ answer stating whether the tests are significant or not if the level of significance is set at 0.05.
- P value summary: The P value asterisk level, if the test was significant.
The beauty of GraphPad Prism is that the interpretation of their statistical analysis is so easy. Obviously, the first thing to look at is the ‘P value‘ column to get the significance value. In all of the tests above, they are above 0.05. Therefore, according to these tests, the data is normally distributed. If it was less than 0.05, then the data will not be normally distributed.
To interpret this more, the ‘Passed normality test (alpha = 0.05)‘ will state is the normality tests have passed, which they have done in this example.
GraphPad Prism version used: 6